Neural Network-based Forecasting Of Sounding Humidity Solar Radiation Error And Query Website Design

At present,more and more digital radiosondes are widely used in atmospheric detection.Compared with the accurate measurement of atmospheric pressure,temperature,wind speed and other high-altitude meteorological elements,the accuracy of humidity measurement needs to be improved.The accuracy error of humidity measurement data is not only from the error of humidity sensor itself,but also easily interfered by other factors in the environment.However,the influence of solar radiation on humidity measurement is often ignored.At present,there is a lack of research in this field.Firstly,this paper studies the phenomenon of radiation and temperature rise of humidity sensor based on the theory of partial drying of solar radiation;Secondly,it uses the fluid dynamics(CFD)software to obtain the data sample set;Then the data set is optimized and compared by neural network algorithm to accurately predict the solar radiation error in humidity measurement;Finally,using Web Storm as the development tool,developing a multifunctional website for high altitude meteorological survey enthusiasts and researchers to inquire,revise and learn by using various front-end and back-end technologies such as vue.js,KOA,Mongo DB,Redis,etc.The specific research contents are as follows:First of all,research is carried out around GTS1 and GTS1-2 radiosondes at home and abroad.In order to prevent the interference of clouds and rain in the high altitude,most of the humidity sensors are equipped with rain caps.Due to the influence of air pressure,solar altitude angle and solar radiation,the temperature around the humidity sensing film of the humidity sensor will be higher than the actual atmospheric temperature due to the air temperature rising inside the rain cap,which will result in the humidity measurement value being dry.Secondly,based on the GTS1-2 measurement system model of Nanjing bridge,the existence of solar radiation drying phenomenon is verified by CFD software.Then,2530 sets of temperature error data samples are simulated by Pro / E modeling,ICEM grid division and FLUENT simulation,taking the typical air pressure,solar altitude angle and solar radiation in the actual high altitude detection as variables.Through the optimization and comparison of BP,PSO-BP,GA-BP and RBF neural network algorithm,RBF neural network algorithm is finally used to build the prediction model,which can predict the solar radiation error of Radiosonde Humidity in different environments,and the predicted value and simulation value are in good agreement.Finally,this article uses Nuxt.js,Vue.js,Koa,Mongo DB,Redis and the other front-end technology full stacks to develop a website for meteorological exploration enthusiasts and researchers to query,revise and learn.The website has the function of user login and registration,users can learn the current knowledge about the detection of high altitude images online;They can check the distribution of the current high altitude image detection stations in various provinces of our country;They can check the weather conditions of various regions;They can calculate the diffusion of solar radiation in various regions and direct radiation;The solar radiation error of sounding humidity can be queried online;They can make webpage corrections to the radiosonde humidity profile.